SAN FRANCISCO & DENVER--(BUSINESS WIRE)--Orange Silicon Valley (OSV), the U.S.-based business innovation center for the global telecommunications provider Orange, and Liqid Inc., the leading provider of composable infrastructure solutions, today unveiled a joint demonstration regarding how next-generation graphics processing (GPU) based supercomputing can be delivered. Powered by Liqid composable infrastructure, the OSV-Liqid prototype is capable of providing on-demand GPU performance as a fully scalable resource. For the first time, users requiring GPU-centric compute can utilize composable platform services, with near-unlimited GPU capacity for the data center’s most computationally demanding applications, including artificial intelligence (AI) and machine learning, virtual and augmented reality (VR/AR), video rendering, DevOps, scientific discovery and other emerging high-value applications. At Supercomputing 2017 in Denver, Liqid is demonstrating the jointly developed GPU configuration live (booth #489).
“To answer the increasing demand for scalability and flexibility for GPU centric infrastructures we envision a new way of introducing agility and composability by extending the low latency PCIe fabric,” said Soumik Sinharoy, Sr. Product Manager, Orange Silicon Valley. “With advanced GPU supercomputing platforms such as these we can meet the demands of emerging applications that benefit from GPU scale out and enable a much wider variety of vertical market segments to take advantage of this powerful, increasingly multi-purpose resource.”
OSV has been a pioneer in GPU-based computing. The company has deployed the resource in its own infrastructure to transcend CPU performance bottlenecks and perform multi-dimensional queries on massive unstructured datasets at speeds far faster than traditional high-performance computing configurations. Simultaneously OSV reduced its infrastructure costs by 40x compared to traditional telecom solutions by embracing GPU for compute.
Liqid’s composable platform enables the scale out and orchestration of GPU alongside NVMe storage, compute and networking, resulting in the ability to efficiently scale core data center resources across PCIe fabric delivering the ideal performance for data-intensive applications that require maximum compute performance.
The OSV and Liqid collaboration enables several industry firsts including:
- Adaptive GPU Supercomputer for artificial intelligence
- GPU scale out - clustering dozens of GPUs across PCIe
- Dynamic, bare-metal orchestration of GPUs to CPU in real time
- Hot-swappable GPU in Linux and Windows
- Peer-to-peer GPU communication via rack-level PCIe fabric
“We believe the future of the data center is composable, with an infinite number of resource allocations to optimize for the world’s most demanding applications,” said Sumit Puri, Vice President, Marketing, Liqid. “Orange Silicon Valley has long recognized that GPUs are becoming central to innovation for next-gen applications in media, business, scientific research and discovery and dev-ops, and we are excited to partner with OSV to increase the user base for GPU-centric, on-demand compute services.”
About Orange Silicon Valley (http://www.orangesv.com/) - Orange Silicon Valley (OSV) is the wholly owned innovation subsidiary of Orange SA, one of the world's leading telecommunications operators, serving 265 million customers across 29 countries. Through research, development, and strategic analysis, Orange Silicon Valley actively participates in the disruptive innovations that are changing the way we communicate. OSV contributes to and engages with the regional Silicon Valley ecosystem through numerous programs, such as our Orange Fab startup accelerator, Orange Institute, and ongoing collaborations with partners. Orange Silicon Valley acts as a guide to the digital revolution occurring in the San Francisco Bay Area, regularly hosting startups, businesses, and corporate leadership from around the world.
A leader in composable infrastructure, Liqid enables users to configure and manage physical, bare-metal server systems in seconds. Storage, compute, networking and graphics processing devices are interconnected over PCI-Express fabric to deliver dynamically configurable bare-metal servers perfectly sized with the exact physical resources required by the application being deployed. Liqid Grid and Command Center allow users to optimize system resources to meet the evolving requirements for next-generation, high-value applications. For more information, contact the Liqid team at firstname.lastname@example.org or visit www.liqid.com. Follow Liqid on Twitter, LinkedIn and Google+.